NYT on October 10 reports Meta Muse agent was launched despite security tests showing password changes and private message access
The New York Times reported on October 10 that internal tests on Meta Muse found the agent could change passwords, read private messages, and leak user addresses to Marketplace sellers without permission, yet CEO Mark Zuckerberg still pushed the September 8 launch.
The New York Times reported on October 10 that Meta internal tests had uncovered a string of serious security and privacy issues in its Muse AI agent, yet CEO Mark Zuckerberg still pushed the product live as originally planned on September 8. According to the report, internal tests showed Muse could change account passwords without user permission, read private messages without authorization and echo them back to the user, and leak private user information such as home addresses to third-party sellers including Facebook Marketplace merchants. Meta Chief AI Officer Wang Tao and AI product lead Nate Friedman were both aware of the risks, but with emerging competitors like Instinct putting pressure on Meta, Zuckerberg chose to ship on schedule. The market response has been strong, with Muse crossing 6.6 million downloads and 1.8 million daily active users within weeks of launch, helped by free pricing and millions of dollars in marketing spend. The episode has reignited industry debate over how to balance development speed and safety in the AI race.